I’m trying to test if there are significant differences between multiple time periods for a group of observations (1 observation per respondent in each time period). For dichotomous outcomes I wanted to use McNemars test but have pweights, so have been using clogit with svy. I think this is legit, but please let me know if I’m forgetting something!

For the categorical variables I want to test (each individually) if the distribution of the proportions across the categories change over time. I was thinking of using signrank, but don’t think that will allow pweights. Any help would be much appreciated!